1 23

Estuaries and CoastsJournal of the Coastal and EstuarineResearch Federation ISSN 1559-2723 Estuaries and CoastsDOI 10.1007/s12237-016-0137-6

Importance and Function of Foraging andRoost Habitat for Wintering AmericanOystercatchers

Janell M. Brush, Amy C. Schwarzer &Peter C. Frederick

1 23

Your article is protected by copyright and

all rights are held exclusively by Coastal

and Estuarine Research Federation (outside

the USA). This e-offprint is for personal

use only and shall not be self-archived

in electronic repositories. If you wish to

self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

Importance and Function of Foraging and Roost Habitatfor Wintering American Oystercatchers

Janell M. Brush1& Amy C. Schwarzer1 & Peter C. Frederick2

Received: 29 October 2015 /Revised: 13 July 2016 /Accepted: 16 July 2016# Coastal and Estuarine Research Federation (outside the USA) 2016

Abstract With changing climate and increased human popu-lations, oyster reefs have been negatively affected by exces-sive wave action; contamination; overharvesting; decreasedfreshwater inputs; and shifts in oxygen, salinity, and turbidityregimes. In Florida’s Big Bend, intertidal reefs dominated bythe eastern oyster (Crassostrea virginica) have experienced anet decline in area of 66 % since the 1980s, a loss likely tohave substantial impacts on reef-dependent wildlife. Ourstudy examined the use of intertidal oyster reefs by winteringAmerican oystercatchers (Haematopus palliatus) in this area.The minimum foraging time required to meet daily caloricneeds was conservatively estimated at 37 min per adult oys-tercatcher, indicating that at present, foraging habitat is not alimiting factor within our study area. We found high-tideroosts to be away from all vegetation and limited in number.They were located in offshore oyster reef habitat, which hasexperienced an 88% decline in area over the past 30 years.Wesuggest that offshore, high-tide roost habitat is a limiting fac-tor and worthy of further attention.

Keywords American oystercatcher .Crassostrea virginica .

Eastern oyster . Estuary .Haematopus palliatus . Oyster reef

Introduction

Rapid changes in marine and estuarine habitats such asseagrass beds, coral reefs, and salt marshes may have wide-ranging implications for the flora and fauna that depend onthese communities. Understanding the specific attributes ofthese habitats that the organisms depend on is key to quanti-fying and projecting effects on different species. Here, weconsider the characteristics of oyster reefs important toroosting and foraging by American oystercatchers(Haematopus palliates) and the susceptibility of these differ-ent reef habitats to ongoing loss.

With an estimated 85 % decline globally, oyster reefcommunities are among the most heavily degraded andendangered marine habitats (Beck et al. 2011). Oystercommuni t i e s domina ted by the eas te rn oys te r(Crassostrea virginica) are highly sensitive to a numberof threats, including excessive wave action; contamina-tion; overharvesting; and shifts in oxygen, salinity, andturbidity regimes (Bahr and Lanier 1981; Beck et al.2009; Seavey et al. 2011; Wall et al. 2005). Elevatedsalinity as a result of decreased freshwater input leadsto increased vulnerability to disease and predation, andfreshwater flow regime has been strongly linked to pro-ductivity and survival of oyster stocks in a number ofestuaries in the USA (Kimmel et al. 2014; Livingstonet al. 1997; Pollack et al. 2011; Wilbur 1992). Althoughthe Florida Big Bend (Crystal River to Panacea; Fig. 1) isa highly conserved and undeveloped part of the GulfCoast (Main and Allen 2007), there has been a 66 %net decline in the extent of the oyster reefs there sincethe 1980s (Baker et al. 2003; Bergquist et al. 2006;Seavey et al. 2011). These reefs are more than 3000 yearsold (Grinnell 1972), which suggests that the rapid declineis a consequence of a recent fundamental change in the

Communicated by James Lovvorn

* Janell M. BrushJanell.Brush@myfwc.com

1 Florida Fish and Wildlife Conservation Commission, 1105 SWWilliston Road, Gainesville, FL 32601, USA

2 Department of Wildlife Ecology and Conservation, University ofFlorida, P.O. Box 110430, Gainesville, FL 32611, USA

Estuaries and CoastsDOI 10.1007/s12237-016-0137-6

Author's personal copy

hydrological conditions that maintain the reefs. Thecauses are probably multiple, but the dominant mecha-nism is increased disease and predation effects resultingfrom high salinities, stemming from reduced freshwaterinput. The increasing length and frequency of low-freshwater discharge events has led to poor recruitmentand survival of oysters, erosion of oyster reefs, and, ulti-mately, irreversible loss of substrate appropriate for oys-ter settlement (Baker et al. 2003; Seavey et al. 2011).

Oyster reefs provide crucial foraging and roosting oppor-tunities for many coast-dependent wintering shorebirds. TheAmerican oystercatcher (hereafter oystercatcher) feeds pri-marily on marine bivalves and depends on coastal areas thatsupport intertidal shellfish beds (Nol and Humphrey 1994).The US population of oystercatchers consists of ca. 11,000individuals (Brown et al. 2005) and is listed as a species ofhigh conservation concern in the US Shorebird ConservationPlan (Brown et al. 2001). Major threats to the oystercatcherinclude widespread habitat loss, increased anthropogenicpressure, and several effects of climate change, particularlysea-level rise (Schulte et al. 2010). Oystercatchers migratefrom northern Atlantic breeding sites to coastal winteringareas on both the South Atlantic US coast and the Gulf ofMexico (Nol and Humphrey 1994). Twomain, geographicallyseparate winter aggregations of the Eastern population exist, atCape Romain, South Carolina (16 % of the estimated total

population size), and in Florida’s Big Bend (8 % of the esti-mated total population size). The Big Bend has been estimatedto support more than 900 individuals from nearly all Atlanticcoastal states during the winter (Schulte et al. 2010). Whileconsiderable research has been focused on breeding habitatconditions, little has been devoted to understanding or im-proving wintering habitat (American Oystercatcher WorkingGroup et al. 2012; Schulte et al. 2010).

The oystercatcher is a long-lived bird, and its populationtrajectories are particularly sensitive to adult survival rates,which may be affected by winter habitat suitability(American Oystercatcher Working Group et al. 2012;Hitchcock and Gratto-Trevor 1997; Yasue 2006). In winter,avian survival is usually sensitive to availability of food, ref-uge from predation, and local weather events (Placyk andHarrington 2004; Sherry and Holmes 1996). Although forag-ing is critical for survival, roosting and its associated activitiessuch as rest, digestion, and maintenance are also important(Conklin et al. 2008). Roost quality is typically associatedwith proximity to feeding habitats because of the energeticcosts of commuting (van Gils et al. 2006). For coastal birds,roost quality is also affected by tidal conditions and protectionfrom strong winds, high surf, precipitation, and predation(Colwell 2010; Cresswell 1994; Gill et al. 2001; Rogerset al. 2006). Shorebirds and seabirds tend to roost in openareas where predators are visible (Clemens et al. 2008).

Fig. 1 Map of study area in theBig Bend of Florida. Theintertidal oyster reef complexincluded in the project is shadedblack

Estuaries and Coasts

Author's personal copy

The rapid decline of oyster reefs in the Big Bend regioncould have significant effects upon a substantial portion ofoystercatcher overwintering habitat. Loss of oyster reefs inthe Big Bend has been disproportionately focused on offshorereefs, with much of the inshore habitat mosaics remaining inrelatively good condition (Seavey et al. 2011). Offshore reefswere defined as those that were more than 1 km from thecoastline and also were not protected from open gulf watersby any other landform. Inshore reefs were defined as thosethat were less than 1 km from the shoreline or were part of themarsh behind the shoreline and were protected from openwater by at least one landform (marsh, sandbar, mudflat, oys-ter reef). Within these mosaics, it is unclear what habitat fea-tures are used most by oystercatchers for feeding and roostingand why.

Based on previous work on oystercatchers, we hypothe-sized that use of foraging habitat would be focused on oysterreefs (a) with high densities of live oysters in a size range thatoystercatchers typically eat (American OystercatcherWorkingGroup et al. 2012; Hand 2008); (b) that were far from vege-tative cover or perches that could harbor aerial or groundpredators (Yasue 2006); (c) that were far from artificial struc-tures (docks, roads, or buildings) that could lead to increaseddisturbance; (d) that had a high edge-to-area ratio that afford ahigh proportion of edge feeding habitat where oysters areinundated (shells open) at water depths accessible to oyster-catchers, which forage by inserting their bills into open oys-ters, tearing the muscle at the hinge to prevent the oysters fromclosing; (e) that had a high density of surrounding foraginghabitat, which would reduce costs of flight among foragingsites; and (f) that were close to roosting sites, which wouldagain reduce costs of flight (van Gils et al. 2006). For roostinghabitat, we expected that oystercatchers would use reefs thatwere (a) high in elevation and large in area, which wouldprovide both shelter from adverse weather conditions and abetter view of predators (Clemens et al. 2008); (b) far fromvegetative cover for mammalian predators and perches foravian predators; (c) distant from artificial structures; (d) farfrom shallow water and close to deep water, which woulddiscourage mammalian predators capable of fording shallowexpanses of water; and (e) close to surrounding foraging hab-itat. We examined oystercatcher use of foraging and roostinghabitat in relation to landscape and microhabitat features tounderstand how current and projected losses of oyster reefsmight affect this declining species.

Materials and Methods

Study Area

The study was conducted in coastal waters from Cedar Key toDeer Island in Levy County, Florida (Fig. 1).While oyster reef

density and coverage vary greatly in the Big Bend region, thestudy area was large enough to include reef systems represen-tative of the area, with largely intertidal reef structure, a gra-dient of inshore marsh to isolated offshore reefs, and a lack ofprotective barrier islands. Offshore reefs were defined as thosethat were more than 1 km from the coastline and also were notprotected from open gulf waters by any other landform.Inshore reefs were defined as those that were less than 1 kmfrom the shoreline or were part of the marsh behind the shore-line and were protected from open water by at least one land-form (marsh, sandbar, mudflat, oyster reef).

Foraging Habitat

We estimated the total number of oystercatchers in the studyarea and identified oyster reefs used by foraging oyster-catchers by performing six ground surveys between October2011 and February 2012. Each survey was conducted by twoor three teams using airboats, covering the entire study area in1 day. Surveys were conducted during tides of varying ampli-tudes within 1.5 h either side of low tide. Teams followed non-overlapping designated routes that allowed 100% coverage ofthe oyster reefs in the survey area. Teams recorded the approx-imate locations (within 75 m) of oystercatchers using a hand-held GPS unit. We also recorded the number of oystercatchers,their behavior (foraging, non-foraging, or both), and the sub-strate type of the reef (sandbar, marsh–oyster, sand–oyster).

In order to characterize the oystercatcher foraging habitat,in ArcMap 10.0 (Esri 2010), we created a fishnet grid(200 × 200 m) on a map of the entire study area. Grid squareswere randomly selected for sampling of foraging habitat. Gridsquares corresponding to locations at which no foraging oys-tercatchers had been observed were eliminated from consid-eration for random selection.

When surveying oyster reefs in the field, the randomlyselected grid squares were sampled for the presence of forag-ing adult oystercatchers. Once detected, the physical locationon the oyster reef where the oystercatcher was foraging (inwater, waterline, top of reef, etc.) was documented. We alsorecorded surface elevation using a laser level from the top ofthe reef and later linked those relative elevations to true ele-vation benchmarks established in the study area. We collectedthe following microhabitat data: slope of the oyster reef at theforaging location, distance to any non-woody emergent vege-tation on the reef (typically short form Spartina alterniflora),and substrate type (mud, sand, shell, mud–shell, sand–shell).Oyster density (living and dead) and presence of musselswithin a 0.25-m2 quadrat were determined 1 m from the oys-tercatcher location, in each cardinal direction. Shell length, thelongest distance from hinge to shell edge (mm), was measuredfor all oysters in one randomly selected quadrat. We also char-acterized nearby foraging habitat by selecting six randompoints within 10 m of the oystercatcher location, where we

Estuaries and Coasts

Author's personal copy

recorded oyster density (living and dead) and presence ofmussels. In every other 0.25-m2 quadrat, we recorded shelllength for all of the oysters within the quadrat.

We also compared characteristics of reefs used for foragingto those of randomly selected reefs to evaluate selection offoraging habitat. We recorded percent cover of oysters, vege-tation, and substrate type on each reef. A laser range finderwas used to measure distance from the center of the reef toboth the nearest woody (typically mangroves or pines) and thenearest non-woody emergent vegetation. Methods used tosample microhabitats depended on the size of the oyster reef.Small oyster reefs (<75 m, longest linear distance) were sam-pled along the long axis of the reef. Large oyster reefs (>75 m,longest linear distance) were sampled along transects perpen-dicular to the long axis of the reef. Randomly generated dis-tances (range 1–5 m) were used to determine the starting pointof sampling as well as each sampling point along each tran-sect. We recorded the following microhabitat informationwithin a 0.50-m2 quadrat at each sample point: percent coverof vegetation, oysters, and open substrate by type (mud, sand,shell, mud–shell, sand–shell); percent living and dead oysters;and presence or absence of mussels. In addition to the fielddata, landscape-level feature data were analyzed inArcMap10.0. These data included area (ha) of the oyster reef,reef shape (0.25 × perimeter)/√area), distance to shallow (0.9–1.8 m) and deep (1.9–3.7 m) waters based on the most recentbathymetric charts, distance (m) to the closest known high-tide and intermediate-tide (available only below mean hightide) roost sites, and density of oyster reefs (m2/ha) within aradius of 250 m.

Foraging Behavior, Prey Selection, and ForagingEnergetics

Foraging adult oystercatchers were observed in order to doc-ument their prey items and estimate average search and han-dling time. From a distance of 15–50 m, we observed adultswithin the same randomly selected grid squares for which wehad made foraging habitat observations. When multiple oys-tercatchers were present, we selected the foraging individualclosest to us in the direction associated with a randomly gen-erated compass degree and observed the bird for 5 min or untilit left the area. Oystercatcher behavior type (e.g., probing,eating, preening) and duration were recorded. The size ofany identified prey item was estimated as a fraction of thelength of the foraging oystercatcher’s bill. The average oys-tercatcher bill length (American Oystercatcher WorkingGroup et al. 2012) was later used to estimate prey length.

In order to estimate the energy contained in the oysters theoystercatchers consumed, we measured mass and length ofoyster meats (longest distance of freshly shucked meat at-tached to bottom valve) of 81 oysters collected from the areaswhere oystercatchers had been seen foraging. Oysters ranged

from 22.4 to 104.5 mm in total shell length. We used thisinformation to develop an equation that explained the relation-ship between fresh meat length and mass of meat (wetmass = 0.2522 e0.0663 (meat length), R2 = 0.91). We then usedequations developed by Dame (1972a) to convert the massesofoystermeats consumed todrymass (drymass=0.1664×wetmass0.97), and from estimated dry mass of meats, we thenestimated total caloric content (5.066 kcal/g dry mass; Dame1972b).

Roosting Habitat

We identified roosts during high and intermediate tides. High-tide roosts were defined as those available to birds at meanhigh tide (Cedar Key gauge) or higher. Intermediate-tideroosts were defined as those available only below mean hightide, typically used at low tide whenwater levels fell below theoyster line and birds ceased foraging. Oystercatchers foragealmost exclusively below the waterline on partially openedoysters presumably because they are easier to manipulateand eat. We located and surveyed all high-tide roosts monthlyfrom August 2011 to February 2012. Intermediate-tide roostswere identified during the six comprehensive ground surveys.Whenever we encountered oystercatchers, we recorded time,weather conditions, and tide stage. In addition, we recordedpercent cover of oysters and vegetation, as well as substratetype on each reef.We also recorded distance from the center ofthe oyster reef to vegetation (both woody and non-woody)using a laser rangefinder, andmaximum elevation was derivedusing survey-grade real-time kinetic GPS equipment correctedwith local benchmarks. We collected the same landscape-levelfeatures (reef area, distance to shallow and deep water, etc.)that we collected for foraging reefs.

Statistical Analyses

In order to characterize oystercatcher foraging habitat, wecompared 40 randomly selected oyster reefs with 40 oysterreefs known to support foraging oystercatchers. We examinedthe effect of the following spatial and physical variables on thepresence of foraging oystercatchers: distance from the reefcenter to nearest woody and non-woody vegetation, distanceto the nearest high-tide roost, distance to shallow and deepwaters, and distance to nearest artificial structure (buildingor road); highest elevation on the reef; percent cover of vege-tation and of oysters; average percentage of oysters that werealive; reef shape; density of surrounding reef (area of adjacentoyster reefs within 250 m); and interaction terms between thevariables.

In order to characterize oystercatcher roosting habitat, wecompared 40 random oyster reefs with 28 intermediate-tideroosts and 13 high-tide roosts. We sampled all the roosts avail-able that had been identified through systematic and

Estuaries and Coasts

Author's personal copy

opportunistic observations. We separated high-tide roostsfrom intermediate-tide roosts because of strong differencesin temporal availability and bird densities at these two roosttypes. The following spatial and physical variables were in-cluded in the analyses: distance from the center of the reef tonearest woody and non-woody vegetation, distance to shallowand deep waters, and distance to nearest structure (building orroad); highest elevation on the reef; percent cover of vegeta-tion; reef area and shape; area of adjacent oyster reefs within250 m; and interaction terms between the variables.

Before analysis, all explanatory variables were tested in apairwise fashion for correlation to preclude issues withmulticollinearity. In cases where variables had a correlationof r ≥ 0.50, one of the variables was excluded from analysis.We used backward stepwise logistic regression (Bewick et al.2005; Hosmer and Lemeshow 2000) in Program R (R 2012)to analyze separately the effect of the variables on use of eachclass of reef (foraging, high tide, intermediate tide). All inde-pendent variables were included in the initial model; then,variables with an alpha level greater than P = 0.10 were re-moved one at a time, starting with the least significant vari-able. Since the habitat we studied was completely open andentirely accessible to the observers via airboat, and becauseoystercatchers are extremely conspicuous in these habitats, wedid not incorporate estimates of detectability in our analyses.

We used Akaike’s information criterion to judge whetherreduced models were an improvement over the full model(Burnham and Anderson 2002). The model with the smallestAkaike’s information criterion (AIC) value was the modelmost strongly supported by the data. If one or more modelswere <2 AIC from the top model, the model with the fewestnumber of parameters was considered to be the most parsimo-nious. The selection of simpler models over more complexmodels when there is no clear difference in explanatory poweris well discussed and has been supported by examples, inwhich simpler models have higher predictive value than morecomplex models (Bolker 2008). All parameters within themost parsimonious models were significant at P ≤ 0.10 unlessotherwise reported.

Results

Foraging Habitat

The number of oystercatchers documented during the six sur-veys ranged from 988 to 1176 (average = 1076; SE = 40.4).We observed the behavior of 163 foraging oystercatchers andcollected microhabitat data on 91 oyster reefs. Average oysterdens i t y a t fo r ag ing s i t e s was 86 .8 oys t e r s /m2

(SE = 2.744 oysters/m2, range 0–660 oysters/m2, n = 1620).Average oyster shell size at foraging sites was 37.8 mm(SE = 0.17 mm, range 2.3–118.0 mm, n = 10,504).

At foraging sites, oysters were arranged in continuousmats, in clumps separated by substrate, or, occasionally, assingle oysters embedded in the substrate. Oystercatchers weremost strongly associated with clumps of oysters (94 % ofobservations), where oyster density is greatest. In addition,oystercatchers foraged almost exclusively (99 % of observa-tions) at or below the waterline, where the oysters are openand easier to access. Oystercatcher foraging areas were com-posed mainly (57 %) of mixed mud and shell.

Based on our predictions, we tested a set of seven candidatemodels to compare reefs used for foraging with randomlyselected reefs (Table 1). The most parsimonious model includ-ed a positive correlation for distance to woody vegetation,negative correlations for distance to nearest high-tide roostand nearest artificial structure, and a positive correlation withaverage percent live oysters (Fig. 2).

Foraging Behavior and Energetics

We observed 62 foraging oystercatchers during two winterseasons. Of the prey items we observed being consumed,72 % were identified. The majority were oysters (95 %), mus-sels accounted for 5 %, and other prey items accounted for<1 %. Many of the unidentified items were probably alsooysters, given the lack of mussels in many parts of the studyarea, but mussels and crabs are common prey items of theoystercatcher in other parts of its range (Hand 2008;Tuckwell and Nol 1997). Average search time, the time be-tween ingestion and locating a new prey item, was 35.2 s(SE = 2.07 s, range 1–202 s); average handling time was11.4 s (SE = ±0.76 s, range 1–50 s, n = 62).

The average bill length in oystercatchers was 86.64 mm(data from multiple North American sites and both sexes;SD = 5.59 mm; American Oystercatcher Working Group et al.2012). Mean oyster meat length in our study averaged 28% ofbill length (SE = 0.01, range 0.1–0.66, n = 144). Using billlength as a guide, this suggests that the average length ofoyster meat consumed was 24.3 mm. We found that the freshmass of 81 local oysters was related to shell length asmass = 0.02552 e0.0663 length. We used this metric to determinethat the average fresh mass of oyster meat consumed by oys-tercatchers in this study was 1.27 g. By equations relating wetto dry mass in oysters found in Dame (1972a) and calorificcontent found in Dame (1972b), the average oyster eaten wasestimated to contain 4.43 kJ. Based on multiples of basalmetabolic rate for different activities, daily energy expenditure(DEE) for breeding male and female oystercatchers has beenestimated at 207 and 226 kJ/14 h/day, respectively (Nol 1985).Given the average of approximately 47 s that an oystercatcherrequired to capture and consume an oyster (search time plushandling time), we calculated that an oystercatcher in ourstudy area would require roughly 37 min of continuous forag-ing per day to satisfy energy needs. This estimate is likely

Estuaries and Coasts

Author's personal copy

high, since the DEE we used was for breeding rather thanindividuals wintering in Florida.

Roosting Habitat

We identified 41 oyster reefs used by oystercatchers duringintermediate (n = 28) and high tides (n = 13).We tested a set of

six models to compare features of random oyster reefs withhigh-tide roosts (Table 2). The most parsimonious of sixmodels comparing random reefs and high-tide roosts includedpositive correlations with elevation, distance to woody vege-tation and area of reef, and a negative correlation with distanceto the nearest artificial structure (Fig. 3). The term for distanceto the nearest artificial structure was insignificant in this

Fig. 2 Probability ofoystercatchers selecting an oysterreef for foraging related to thevariables included in the mostparsimonious model (Table 1).The solid lines indicate themodel-based predictions for eachvariable with all other variablesheld at their mean values. Thedashed lines indicate the 95 %confidence intervals. All variablesare plotted from the 5th to 95thpercentiles of the observed values

Table 1 Number of parameters (#Par), AICc values, and log-likelihood (LL) for candidate models comparing foraging reefs with randomly selectedreefs

Model #Par AICc ΔAICc AICc weight LL

Dist_Woody + Dist_HiRoost + Dist_Structure + avg_live 5 100.45 0 0.31 −44.82Dist_Woody + Shape + Dist_HiRoost + Dist_Structure + avg_live 6 100.55 0.1 0.29 −43.69Dist_Woody + Shape + Shallow + Dist_HiRoost + Dist_Structure + 7 101.28 0.83 0.2 −42.85Dist_Woody + Shape + Shallow + Dist_HiRoost + Dist_Structure +

Dist_Non-Woody + avg_live8 101.97 1.52 0.14 −41.96

Dist_Woody + Shape + Shallow + Dist_HiRoost + Dist_Structure +Dist_Non-Woody + Dist_Non-Woody × Dist_Woody + avg_live

9 104.39 3.94 0.04 −41.89

Dist_Woody + Total_Veg + Shape + Shallow + Dist_HiRoost +Dist_Structure + Dist_Non-Woody + Dist_Non-Woody ×Dist_Woody + avg_live

10 106.88 6.42 0.01 −41.82

Dist_Woody + Total_Veg + %_oys + Shape + Shallow + Density +Dist_HiRoost + Dist_Structure + Dist_Non-Woody + Dist_Non-Woody ×Dist_Woody + avg_live

12 112.34 11.88 0 −41.8

Variables include distance to woody vegetation (Dist_Woody), distance to non-woody vegetation (Dist_Non-Woody), distance to high-tide roost (Dist_HiRoost), distance to artificial structure (Dist_Structure), distance to shallow water (Shallow), percent cover of oysters (%_oys), average number of liveoysters/m2 (avg_live), percent cover of all vegetation (Total_Veg), shape (Shape), and density of oyster reefs (m2 /ha) within 250 m (density)

Estuaries and Coasts

Author's personal copy

model; however, removal of this term resulted in a model fit>2 AIC below the most parsimonious model. The small sam-ple size of the high-tide roosts resulted in large confidenceintervals in the results.

We tested a set of four models to compare features of ran-dom oyster reefs with intermediate roosts (Table 3). The mostparsimonious model included positive correlations for dis-tance to woody vegetation, non-woody vegetation, and shal-low water. Intermediate-tide roost habitat was negatively cor-related with density of oyster reefs within 250 m and positive-ly correlated to the interaction term between distance to

woody vegetation and distance to non-woody vegetation(Fig. 4).

Discussion

The average prey search and handling times of wintering adultoystercatchers were very similar to those observed in the CapeRomain region of South Carolina (Hand et al. 2010; Sanderset al. 2013), where the largest US population of winteringAmerican oystercatchers resides. Our estimates indicated that

Table 2 Number of parameters (#Par), AICc values, and log-likelihood (LL) for the models comparing high-tide roosts with randomly selected reefs

Model #Par AICc ΔAICc AICc weight LL

Elevation + Dist_Woody + Area + Dist_Structure 5 23.29 0 0.53 −6.01Elevation + Dist_Woody + Dist_Non-Woody + Area +

Dist_Structure6 24.71 1.41 0.26 −5.44

Elevation + Dist_Woody + Dist_Non-Woody + Area +Shallow + Dist_Structure

7 26.3 3.01 0.12 −4.91

Elevation + Dist_Woody + Area 4 27.87 4.58 0.05 −9.52Elevation + Dist_Woody + Dist_Non-Woody + Area +

Shallow + Deep + Density + Dist_Structure9 28.91 5.61 0.03 −3.36

Elevation + Dist_Woody + Dist_Non-Woody + Total_Veg +Area+ Shallow + Deep + Density + Dist_Structure

10 31.96 8.66 0.01 −3.36

Variables include distance to woody vegetation (Dist_Woody), distance to non-woody vegetation (Dist_Non-Woody), distance to artificial structure(Dist_Structure), distance to shallow water (Shallow) and deep water (Deep), maximum elevation (Elevation), percent cover of all vegetation (Total_Veg), reef area (Area), and density of oyster reefs (m2 /ha) within 250 m (Density)

Fig. 3 The probability ofoystercatchers selecting an oysterreef as a high-tide roosting loca-tion related to the habitat variablesincluded in the most parsimoni-ous model (Table 2). Elevation inthe first graph is relative to meansea level. The solid lines indicatethe model-based predictions foreach variable with all other vari-ables held at their mean values.The dashed lines indicate the95 % confidence intervals. Allvariables are plotted from the 5thto 95th percentiles of the observedvalues

Estuaries and Coasts

Author's personal copy

an oystercatcher would need to forage for 37 min daily tosatisfy energy requirements in Florida. Since this estimatewas based on daily energy requirements during the breedingseason, we believe it to be a liberal estimate of winter energyrequirements. This finding suggests that prey availability isnot a limiting factor for this wintering population.Nevertheless, most of the foraging areas used by oyster-catchers were on the nearshore and inshore reefs; many ofwhich are sinking or exhibiting low accretion rates (Seaveyet al. 2011). Given the rapid loss of reef area in the region(>66 % in 30 years), this preferred foraging habitat may soonbe at risk (Seavey et al. 2011), limiting foraging habitat for thiswintering population.

Intermediate-tide roosts were used only when the waterlevel had dropped below that of the oysters on the reefs usedfor foraging. Since intermediate-tide roosts were interspersedamong foraging areas, the threats facing this complex of reefsare probably similar to those facing foraging reefs. Oysterreefs selected as intermediate-tide roosts differed in importantways from reefs selected for foraging. Foraging oystercatcherstolerated non-woody vegetation, but roosting oystercatcherspreferred roosts without, and distant from, non-woody vege-tation. Intermediate-tide roosts were also located a greater dis-tance from shallow water, although this effect was weak(Fig. 4). This preference may relate to avoidance of terrestrialpredators capable of swimming short distances. All of thesefeatures suggest that intermediate-tide roosts tend to be fartherfrom the coastline than random reefs. Intermediate-tide roosts,however, were still located within the matrix of nearshorereefs, unlike high-tide roosts, which were located offshore.Nearshore reefs are less likely than offshore reefs to be lostunder current conditions (Seavey et al. 2011).

High-tide roosts in the study area were used both day andnight and have a critical function for the oystercatchers.Although high-tide roosts were significantly more elevatedthan randomly selected reefs, many are already threatenedby overwash and erosion during the normal tidal cycle.Between 1982 and 2011, Florida’s Big Bend had an average

of 66% net loss of oyster reef area. Offshore reefs had an evenhigher rate of loss at 88 % (Seavey et al. 2011). Eight of 13high-tide roosts and all but one of the roosts used at night arelocated on offshore reefs, indicating that in the study area,high-tide roosting habitat is likely at greater risk of declinethan is intermediate-tide roost habitat. Given the small numberof available roosts (13 high-tide roosts versus 28 intermediate-tide roosts) and their immediate vulnerability, the amount ofhigh-tide roosting habitat is likely to be a limiting factor forthis population. In addition, there are no areas on the GulfCoast of Florida that resemble the wintering habitat used byoystercatchers in Cedar Key, making it impossible to findsimilar habitat elsewhere if the habitat in Cedar Key becomescompromised.

Our analysis suggests that oystercatchers use roosting sitesthat are far from or lacking woody vegetation; this may beattributed to woody vegetation’s providing cover for mamma-lian predators (e.g., mink, raccoons) and perches for raptors(Yasue 2006). Predation by raptors is an important source ofadult mortality (American Oystercatcher Recovery WorkingGroup 2012). Peregrine falcons (Falco peregrinus) and baldeagles (Haliaeetus leucocephalus) have frequently used someof these preferred roosts. If oystercatchers are forced to usesuboptimal roosting locations, there could be survival andenergy costs associated with increased predation and in-creased antipredator behaviors (Rogers et al. 2006).Oystercatchers also face further competition for optimalhigh-tide roost sites as human recreational use of these fewexposed reefs at high tide continues to increase. Florida hasdocumented a 20 % increase in tourism over the past 10 years(VISIT FLORIDA research 2016), and many of these touristsuse marine resources.

We suggest that in the Big Bend, oystercatchers are forcedto choose between rapidly declining optimal roost habitat inoffshore areas and more abundant but suboptimal high-tideroost habitat closer to shore. Oyster reef habitat in the regionof Cape Romain region is also declining (Sanders et al. 2004).Oyster reef habitat is more extensive in Cape Romain than it is

Table 3 Number of parameters (#Par), AICc values, and log-likelihood (LL) for the models comparing intermediate roosts with randomly selectedreefs

Model #Par AICc ΔAICc AICc weight LL

Dist_Woody + Dist_Non-Woody + Area + Shallow + Density +Dist_Non-Woody × Dist_Woody

7 71.79 0 0.34 −27.96

Dist_Woody + Dist_Non-Woody + Shallow + Density +Dist_Non-Woody × Dist_Woody

6 72.01 0.22 0.31 −29.31

Dist_Woody + Dist_Non-Woody + Total_Veg + Area+ Shallow +Density + Dist_Non-Woody × Dist_Woody

8 72.24 0.45 0.27 −26.9

Dist_Woody + Dist_Non-Woody + Total_Veg + Area+ Shallow +Density + Dist_Structure + Dist_Non-Woody × Dist_Woody

9 74.65 2.86 0.08 −26.77

Variables include distance to woody vegetation (Dist_Woody), distance to non-woody vegetation (Dist_Non-Woody), distance to artificial structure(Dist_Sturcture), distance to shallow water (Shallow) and deep water (Deep), maximum elevation (Elevation), percent cover of all vegetation (Total_Veg), reef area (Area), and density of oyster reefs (m2 /ha) within 250 m (Density)

Estuaries and Coasts

Author's personal copy

in the Big Bend of Florida, and while Cape Romain alsoaccommodates a larger population of wintering oyster-catchers, existing high-tide roost habitat is more than suffi-cient for that population’s needs.

Since oyster reefs can grow vertically much faster than sealevel is rising (Rodriguez et al. 2014), the loss of habitat isprobably not due primarily to sea-level rise. Ultimately, theloss of oyster populations is linked to reductions in freshwaterinputs to estuaries, a result of changing freshwater

management upstream (Seavey et al. 2011). While effortsare being made to conserve upstream water resources andreduce anthropogenic impacts on frequency or severity oflow-flow events, these policies may take decades to enact. Inthis context, an effective conservation strategy is to minimizethe likelihood of further losses in oyster reefs used as high-tideroosting habitat for oystercatchers by adding durable substrateto the degraded reefs. This is likely to allow repeated recolo-nization of reefs by oysters following episodic die-offs. Thereis no apparent shortage of larvae in this nearshore ecosystem(Frederick and Sturmer, unpublished data), probably becausemainland marsh or tidal creek oyster populations have notdeclined as much as offshore reefs (Seavey et al. 2011). Theexistence of durable substrate would allow oysters to repeat-edly recolonize reefs, making them resilient in the long term tofluctuations in freshwater flow.

Acknowledgments This work was funded by the National Fish andWildlife Federation Shell Grant Program.We wish to thank the followinggroups: US Fish and Wildlife Service, American Oystercatcher WorkingGroup, University of Florida Institute of Food and Agricultural Science,USGS Florida Cooperative Fish and Wildlife Research Unit, USGSNorth Carolina Cooperative Fish and Wildlife Research Unit, andLower Suwannee and Cedar Keys National Wildlife Refuge. Thank youto the many individuals who contributed to the success of this project, inparticular Bobbi Carpenter, Andrew Cox, Carolyn Enloe, Andrew Gude,Doris Leary, Pat Leary, Erin Leone, Katharine Malachowski, JennSeavey, Nick Vitale, and Jeremy Wood.

References

American Oystercatcher Working Group, Nol, Erica, and Humphrey,Robert C. 2012. American oystercatcher (Haematopus palliatus),The birds of North America online (A. Poole, ed.). Ithaca, NewYork: Cornell University, Cornell Lab of Ornithology. http://bna.birds.cornell.edu/bna/species/082. Accessed 01 Dec 2015.

Bahr, L.M., and Lanier W.P.. 1981. The ecology of intertidal oyster reefsof the South Atlantic coast: a community profile. FWS/OBS-81/15.Washington, D.C.: U.S. Fish and Wildlife Service, Office ofBiological Services.

Baker, P., Bergquist D., and Baker S.. 2003.Oyster reef assessment in theSuwannee River estuary. Final report. Jacksonville, Florida: St.Johns River Water Management District.

Beck, M., R. Brumbaugh, L. Airoldi, A. Carranza, L. Coen, C. Crawford,O. Defeo, G. Edgar, B. Hancock, M. Kay, H. Lenihan, M.W.Luckenbach, C.L. Toropova, and G. Zhang. 2009. Shellfish reefsat risk: a global analysis of problems and solutions. Arlington,Virginia: The Nature Conservancy.

Beck, M.W., R.D. Brumbaugh, L. Airoldi, A. Carranza, L.D. Coen,C. Crawford, O. Defeo, T.J. Edgar, B. Hancock, M.C. Kay,H.S. Lenihan, M.W. Luckenbach, C.L. Toropova, G.F. Zhang,and X.M. Guo. 2011. Oyster reefs at risk and recommendationsfor conservation, restoration, and management. Bioscience 61:107–116.

Bergquist, D.C., J.A. Hale, P. Baker, and S.M. Baker. 2006. Developmentof ecosystem indicators for the Suwannee River estuary: oyster reefhabitat quality along a salinity gradient. Estuaries and Coasts 29:353–360.

Fig. 4 The probability of oystercatchers selecting an oyster reef as anintermediate-tide roosting location related to the habitat variables includ-ed in the most parsimonious model (Table 3). The solid lines indicate themodel-based predictions for each variable with all other variables held attheir mean values. The dashed lines indicate the 95 % confidence inter-vals. All variables are plotted from the 5th to 95th percentiles of theobserved values. In the graph representing the interaction term, the linesrepresent the effect size of distance from non-woody vegetation at givendistances fromwoody vegetation. Confidence intervals were excluded forvisual clarity

Estuaries and Coasts

Author's personal copy

Bewick, V., L. Cheek, and J. Ball. 2005. Statistics review 14: logisticregression. Critical Care 9(1): 112–118.

Bolker, Benjamin M. 2008. Ecological models and data in R. Princeton:Princeton University Press.

Brown, S., C. Hickey, B. Harrington, and R. Gill, eds. 2001. The U.S.shorebird conservation plan, 2nd edn. Manomet, Massachusetts:Manomet Center for Conservation Sciences.

Brown, S., S. Schulte, B. Harrington, B. Winn, J. Bart, and M. Howe.2005. Population size and winter distribution of eastern Americanoystercatchers. Journal of Wildlife Management 69: 1538–1545.

Burnham, K.P., and D.R. Anderson. 2002. Model selection andmultimodel inference: a practical information-theoretic approach,2nd edn. New York: Springer-Verlag.

Clemens, R.S., Haslem A., Oldland J., Shelley L., Weston M.A., andDiyan M.A.A.. 2008. Identification of significant shorebird areasin Australia: mapping, thresholds and criteria. Birds Australia reportfor the Australian Government’s Department of Environment andWater Resources.

Colwell, M.A. 2010. Shorebird ecology, conservation, and management.Berkeley, California: University of California Press.

Conklin, J.R., M.A. Colwell, and N.W. Fox-Fernandez. 2008. High var-iation in roost use by dunlin wintering in California: implications forhabitat limitation. Bird Conservation International 18: 275–291.

Cresswell, W. 1994. Age-dependent choice of redshank (Tringa totanus)feeding location: profitability or risk? Journal of Animal Ecology63(3): 589–600.

Dame, R.F. 1972a. Comparison of various allometric relationships inintertidal and subtidal American oysters. Fishery Bulletin 70(4):1121–1126.

Dame, R.F. 1972b. Variations in the caloric values of the American oysterCrassostrea virginica. Proceedings of the National ShellfisheriesAssociation 62: 86–88.

Gill, J.A., K. Norris, and W.J. Sutherland. 2001. The effects of distur-bance on habitat use by black-tailed godwit Limosa. Journal ofApplied Ecology 38: 846–856.

Grinnell, R.S. Jr. 1972. Structure and development of oyster reefs on theSuwannee River delta, Florida. Ph.D. dissertation, State Universityof New York, Binghamton.

Hand, C. 2008. Foraging ecology of American oystercatchers in the CapeRomain region, South Carolina. M.S. thesis, Clemson University,Clemson, South Carolina.

Hand, C.E., F.J. Sanders, and P.G.R. Jodice. 2010. Foraging proficiencyduring the nonbreeding season of a specialized forager: are juvenileAmerican oystercatchers Bbumble-beaks^ compared to adults?Condor 112(4): 670–675.

Hitchcock, C.L., and C. Gratto-Trevor. 1997. Diagnosing a shorebirdlocal population decline with a stage-structured population model.Ecology 78: 522–534.

Hosmer, D.W., and S. Lemeshow. 2000. Applied logistic regression. NewYork: Wiley and Sons.

Kimmel, D.G., M. Tarnowski, and R.I.E. Newell. 2014. The relationshipbetween interannual climate variability and juvenile eastern oysterabundance at a regional scale in Chesapeake Bay. North AmericanJournal of Fisheries Management 34: 1–15.

Livingston, R.J., X. Niu, F.G. Lewis III, and G.C. Woodsum. 1997.Freshwater input to a gulf estuary: long-term control of trophic or-ganization. Ecological Applications 7: 277–299.

Main, M.B., and G.M. Allen. 2007. Florida’s environment: north centralregion. Gainesville, Florida: Wildlife Ecology and Conservation

Department, Florida Cooperative Extension Service, Institute ofFood and Agricultural Sciences, University of Florida.

Nol, E. 1985. Sex roles in the American oystercatcher. Behaviour 95:232–260.

Nol, E., and R. Humphrey. 1994. American oystercatcher (Haematopuspalliatus). In The birds of North America online, ed. A. Poole.Ithaca, New York: Cornell Lab of Ornithology, Cornell University.

Placyk, J.S., and B.A. Harrington. 2004. Prey abundance and habitat useby migratory shorebirds at coastal stopover sites in Connecticut.Journal of Field Ornithology 75(3): 223–231.

Pollack, J.B., H. Kim, E.K. Morgan, and P.A. Montagna. 2011. Role offlood disturbance in natural oyster (Crassostrea virginica) popula-tion maintenance in an estuary in South Texas, USA. Estuaries andCoasts 34: 187–197.

Rodriguez, A.B., F.J. Fodrie, J.T. Ridge, N.L. Lindquist, E.J. Theurkauf,S.E. Coleman, J.H. Grabowski, M.C. Brodeur, R.K. Gittman, D.A.Keller, and M.D. Kenworthy. 2014. Oyster reefs can outpace sea-level rise. Nature Climate Change 4: 493–497.

Rogers, D., T. Piersma, and C. Hassell. 2006. Roost availability mayconstrain shorebird distribution: exploring the energetic costs ofroosting and disturbance around a tropical bay. BiologicalConservation 133: 225–235.

Sanders, F.J., T.M.Murphy, andM.D. Spinks. 2004.Winter abundance ofthe American oystercatcher in South Carolina. Waterbirds 27(1):83–88.

Sanders, F., M. Spinks, and T. Magarian. 2013. American oystercatcherwinter roosting and foraging ecology at Cape Romain, SouthCarolina.Wader Study Group Bulletin 120(2): 128–133.

Schulte, S., Brown S., Reynolds, D. and the American OystercatcherWorking Group. 2010. American oystercatcher conservation actionplan for the United States Atlantic and Gulf coasts, Version 2.1.

Seavey, J.R., W.E. Pine III, P. Frederick, L. Sturmer, and M. Berrigan.2011. Decadal changes in oyster reefs in the big bend of Florida’sgulf coast. Ecosphere 2(10): 114.

Sherry, T.W., and R.T. Holmes. 1996. Winter habitat quality, populationlimitation, and conservation of Neotropical–Nearctic migrant birds.Ecology 77: 36–48.

Tuckwell, J., and E. Nol. 1997. Foraging behavior of American oyster-catchers in response to declining prey densities. Canadian Journalof Zoology 75: 170–181.

van Gils, J.A., T. Piersma, A. Dekinga, B. Spanns, and C. Kraan. 2006.Shellfish dredging pushes a flexible avian top predator out of amarine protected area. PLoS Biology 4(12): e376. doi:10.1371/journal.pbio.0040376.

VISIT FLORIDA research. Historic visitor estimates. http://www.visitfloridamediablog.com/home/florida-facts/research/. Accessed11 March 2016.

Wall, L., L. Walters, P. Sacks, and R. Grizzle. 2005. Recreational boatingactivity and its impact on the recruitment and survival of the oyster(Crassostrea virginica) on intertidal reefs in mosquito lagoon,Florida. Journal of Shellfish Research 24(4): 965–974.

Wilbur, D. 1992. Associations between freshwater inflows and oysterproductivity in Apalachicola Bay, Florida. Estuarine, Coastal andShelf Science 35: 179–190.

Yasue, M. 2006. Environmental factors and spatial scale influence shore-birds’ responses to human disturbance. Biological Conservation128: 47–54.

Estuaries and Coasts

Author's personal copy